Folks, I added a new ticket with the newcomer label for cucumber-ruby: https://github.com/cucumber/cucumber-ruby/issues/1137 I’m worried I didn’t add enough explanation (I was in a bit of a rush) what do you think?
Ok here goes my first PR https://github.com/cucumber/cucumber-ruby/pull/1140
*Thread Reply:* @aida Finding ways to reduce
"Assignment Branch Condition size" when the score is missing out 2.4 points : [102.4/100]
is more GOLF than logic.
*Thread Reply:* yeah, when I saw I was like f** haha, but I think I found a way
*Thread Reply:* In this case: # rubocop:disable Metrics/AbcSize placed on the def parse! line as a suffix comment would shut this up.
*Thread Reply:* Any option parser will be tough to keep a small AbcSize on.
*Thread Reply:* @aida some of that is my fault - I added a script this morning that fails rubocop!
*Thread Reply:* @aida if you rebase now rubocop should be happier
*Thread Reply:* Sorry, I should check: do you know what I mean by “rebase”?
*Thread Reply:* I believe the build should be green as of 9ad46cb4ca122a7ac776c6de64331cdb3be67842
*Thread Reply:* but I’m not sure what’s going on with the Windows / AppVeyor build just yet
hi there! I'm working in the PR https://github.com/cucumber/cucumber-ruby/pull/1132 about the issue https://github.com/cucumber/cucumber-ruby/issues/1113
*Thread Reply:* Ei! @matt-SmartBear how was your vacation? one of the contributors modified the test scenario and the PR was merged
*Thread Reply:* Woot! How was the experience @madame.sheema? Did we treat you well? Are you keen to do more or did the experience put you off?
*Thread Reply:* @matt-SmartBear the experience was really good, and for sure that I'm going to continue doing things, but for now I will try to pick up easy ones 😊
@Gem a good place to get started contributing to any Cucumber repo is to look for the “newcomers” label. We’re still getting things tidied up to add those labels in the right places, but it’s a start:
https://github.com/cucumber/cucumber-ruby/issues?q=is%3Aopen+is%3Aissue+label%3Anewcomers
*Thread Reply:* @matt-SmartBear Perhaps it's my not knowing multiple ways to set things up appropriately, but things like this line or the :codeswarm task below are examples https://github.com/cucumber/cucumber-ruby/blob/master/gem_tasks/contributors.rake#L6
*Thread Reply:* How about this line? https://github.com/cucumber/cucumber-ruby/blob/master/lib/autotest/cucumber_mixin.rb#L114
*Thread Reply:* Nearly this whole file 😛 https://github.com/cucumber/cucumber-ruby/blob/master/lib/cucumber/cli/options.rb
*Thread Reply:* Even in my own Cucumber files, I've got step defs that hit 90 characters. Should I be refactoring these to be shorter, too?
@matt-SmartBear On PR 1173 about the duplicate methods: It seems that inlining them may not be as simple as initially thought. Each implementation I've tried, I end up getting different ordered data back so tests fail. Seeing as you mentioned about asking any questions, in here, I'm curious if you've got some insight into that area of the code base?
I've cleaned up some messages as I'm getting more of an understanding of these associated files. Not completely, but more! haha. It's curious that in the entirety of the repo this is the only line that calls MultilineArgument.doc_string directly. https://github.com/jaysonesmith/cucumber-ruby/blob/1021-fix-style-violations-2/lib/cucumber/runtime/support_code.rb#L37
*Thread Reply:* I propose that all of lines 35-43 go into MultilineArgument.from if possible.
I’m sorry I’m too rusty on that bit of the codebase to be able to give specific advice, but that clump there looks like it’s begging to be extracted into a factory method.
*Thread Reply:* Let me try to rework that section to get that going. I ended up pushing a different solution to the PR, but it's definitely not the same thing as what you're requesting.
*Thread Reply:* @matt-SmartBear Give it a look-see now? https://github.com/cucumber/cucumber-ruby/pull/1173/commits/0af9db95ce61a68c42aefc130f852ea8d4457b8a
*Thread Reply:* That would be really useful. I think this belongs in contributing.md
@Gem this is the code for the @Aslak Hellesøy bot that invites new contributors: https://github.com/cucumber/commitbit
The welcome message source is here: https://github.com/cucumber/commitbit/blob/master/welcome.ejs.template
Bear in mind that the same message is used across all repos (at the moment) so we’d need to keep it generic enough to make sense in all contexts. I guess pointing to the repo’s contributing.md would be OK though, they should all have one of those.
I’m not quite sure how changes are deployed. Manually at the moment I think.
*Thread Reply:* > I’m not quite sure how changes are deployed. Manually at the moment I think.
As described here @matt-SmartBear: https://github.com/cucumber/commitbit#deploy-the-service
*Thread Reply:* is that just for cucumber ltd folks or core team or what?
*Thread Reply:* sorry if these questions are dumb, I’m still way behind on email since cukenfest
*Thread Reply:* More clear now @matt-SmartBear? https://github.com/cucumber/commitbit#deploy-the-service
*Thread Reply:* yeah better. who exactly could I contact? does it have to be someone in cucumber ltd?
*Thread Reply:* I don't expect tons of changes to commit bit though 🙂
*Thread Reply:* I can see us building other bots too so it might grow
*Thread Reply:* I’m just thinking about @Gem - he has feedback on the template, so the docs need to be self-explanatory for him.
*Thread Reply:* A new bot should probably be a new microservice. I highly recommend playing with Serverless framework and AWS lambda. It's awesome!
*Thread Reply:* Nope, just some node stuff. One instance I use just kicks off a CircleCI rebuild that then runs some cucumber tests that drive a handful of devices through SauceLabs.
*Thread Reply:* The other is a series of dummy advertising responses also in Node. I handle a lot of different query strings and return back generated xml
*Thread Reply:* it just formats the line in a consistent way based on the PR content, then inserts it at the top of the right section
*Thread Reply:* Its big win is formatting some additional texxxxt - what you said
*Thread Reply:* the other thing you want to do when merging other people’s is check that the PR’s title makes sense
*Thread Reply:* If you just said, 'go update this' I imagine people would have their own idea about the color of the bike shed.
*Thread Reply:* Well @brasmusson feels strongly that the new entries should go at the top, and I kept absent-mindedly adding them at the bottom and getting on his nerves 🙂
*Thread Reply:* Is this the correct placement of the new additions? I feel like it should be under the 3.0.0.pre.2 section?
*Thread Reply:* @Gem 3.0.0.pre.2 has been released so it’s too late to get any changes into that
*Thread Reply:* Ah okay. Should there be a new version number added to the history file then or just leave it be for now?
*Thread Reply:* To be fair, we should have updated lib/cucumber/version when 3.0.0.pre.2 was released with the next target release for master
*Thread Reply:* TL;DR I value consistency (a lot). Yes, I prefer to organize changelogs according to the "latest change first" policy.
When I start to make changes in a code file, I really try to follow the code conversions used in that file. For instance if the existing variable names are using camel-case, any variables I add will also get camel-case names. My personal preference for variable name style is irrelevant, consistency is much more important.
The same goes for changelogs/history-files. I'd like to be able to use "git blame" and with a glance see what policy is to be used in the project, and then follow that policy - to maintain consistency. My personal preference for changelogs/history-files is to use the "latest changes first" policy, both with regards to ordering releases and with regards to ordering items within (sub sections within) each release. I'm not happy if the changelog/history-file has version 1 at the top, and I have to scroll down forever to set what has happend lately (at work - in documents - I have to live with it, having to browse several pages of the amendment record to see what changes have been made in the latest revision, my comment to that is "seriously? seriously?).
The problem for my with @matt-SmartBear adding entries to the bottom of sections in the history-file, is that it raises doubts in my mind that there either is a policy I do not know about, or (worse) that there is no shared understanding among us contributors on the editing of the changelog/history file. Either way that is a stumbling block that makes the ½hour or so I have to just merge the PR at hand and be done with it is not enough time. Before doing that I have to find out whether there is a policy I do not know about, or work on creating a shared understanding - so I then can update the changelog/history-file with consistency. This all ends up that I have to do the work at hand (merging a PR that is ready to be merged, for instance) later - and we all know that "later means never".
If there is not consistency in the changelog/history-file, then a task that should be quite straight forward, suddenly is much harder (at least for me).
While we’re talking about new additions by newish people; https://github.com/cucumber/cucumber-ruby/pull/1153 I’m going to go ahead and merge that.
*Thread Reply:* consistency / conventions are good with me. Is there anything established from previous commits?
*Thread Reply:* They're basically on par with my latter example, so nothing too structured aside from "Fix Foo"
*Thread Reply:* I’m sure starting a convention of prefixing wth “Robocop” can’t do any harm
*Thread Reply:* @matt-SmartBear Basically anything to do with the cli output is a good example. Really long strings for errors or config tables.
*Thread Reply:* I'd have to run rubocop to get more of the non-string/output examples. Not on my osx dev machine currently
*Thread Reply:* @Gem I think the 80 character limit would be useful to apply here. You could miss those conditionals right over at the end of the line.
*Thread Reply:* You could make specific error types (subclassing ProfileNotFound, YmlLoadError) that own the responsibility of constructing their error message.
*Thread Reply:* That would keep the logic and the presentation separate, which would be neater.
*Thread Reply:* If the lines in the new error types are too long, try using something like this:
def message
[
'line 1',
'line 2'
].join
end
*Thread Reply:* Thinking about it, if these lines are over 80 characters in our editors, they’ll also be 80 characters in our users cmd prompts
*Thread Reply:* So maybe we should think about that and just format them nicely over multiple lines, and use a heredoc instead. (https://infinum.co/the-capsized-eight/multiline-strings-ruby-2-3-0-the-squiggly-heredoc)
If anyone’s interested, this little fix could do with some love: https://github.com/cucumber/cucumber-ruby/pull/919
*Thread Reply:* on that, I think that the problem with people updating the history file in their PR is that it will easily create merge conflicts.
*Thread Reply:* Oh, I wasn't advocating the act of updating history files as well in their PRs.
*Thread Reply:* Was simply noting a difference in the information included in the c++ contrib file versus the ruby contrib file. 🙂
*Thread Reply:* @theo has been shorting that form yesterday to try and make it more friendly. It’s really just the very first part of a conversation with a company about providing training, and we don’t need to know programming language at that early stage.
*Thread Reply:* Perhaps the wording “have it set in your environment to GITHUB_TOKEN” may confuse a reader.
“and set the environment variable GITHUB_TOKEN to the token’s value” (Gah, is that even more confusing?)
*Thread Reply:* @matt-SmartBear I'd rather have it as part of the new website docs, then individual repos' CONTRIBUTING.md can link to that.
*Thread Reply:* We need something in those contributing files though, no? Look what’s there at the moment: https://github.com/cucumber/cucumber/blob/master/CONTRIBUTING.md
*Thread Reply:* I agree some CONTRIBUTING.md files need to be updated, but I'd rather keep them all slim, linking to an easy-to-read doc on the docs site.
*Thread Reply:* We can link in to the monorepo’s contributing file from the individual repos’ ones.
*Thread Reply:* I agree, ultimately that would be great, but why don’t we iterate towards that.
*Thread Reply:* If we move your content out of the mailing list post into that file, it will already be easier for people (like @Gem) to find.
*Thread Reply:* We can, but I'd rather have something like this: http://hood.ie/contribute/ - with some illustrations.
*Thread Reply:* I worry that’s making the perfect the enemy of the good
*Thread Reply:* for js, they just tag the commit and the build server does it for them!
*Thread Reply:* Taking a quick look at the cucumber-js contributing versus cucumber-ruby, they are both a good bit different
*Thread Reply:* I'd love to settle on the same format for changelogs across the org. I suggest http://keepachangelog.com/
@Aslak Hellesøy @matt-SmartBear Just finished writing this up: https://github.com/cucumber/cucumber/issues/251
I just added a newcomers ticket for cucumber-ruby folks: https://github.com/cucumber/cucumber-ruby/issues/1198
Hello. Re: https://github.com/cucumber/cucumber-jvm/pull/1025, coincidentally, we are now planning on working on this again. We had other priorities on our end in the past, but now we are focusing on testing strategy again and need to get this resolved. So we will be working on this soon. Please be patient with us.
*Thread Reply:* So happy to see you are doing this again!! Unfortunately I won’t be able to join this weekend
Want to contribute to Cucumber? Here is a little issue to get you started: https://cucumberbdd.slack.com/archives/C5WD8SA21/p1534891342000100
*Thread Reply:* Had any help with this yet? I'm into helping with the javascript and ruby (I've never touched go).
*Thread Reply:* Actually, the go lang should be easy too. Looks like just moving the logic you've already worked out in Java over to
> .../javascript/.../tree_regexp.js
> .../ruby/.../tree_regexp.rb
> .../go/tree_regexp.go
This really is a great one to get started on.
*Thread Reply:* Sounds good! It’s almost like painting by numbers. You can hack on a language without knowing it, because you’ve seen the implementation in a different language.
*Thread Reply:* @mikegfisher the reporter said he’d work on it tonight: https://github.com/cucumber/cucumber/issues/454#issuecomment-414854132 - maybe sync up with him?
<a href="https://github.com/cucumber/cucumber">cucumber/cucumber</a>
Hi Team.... This is Raja... I am planning to integrate our cucumber automation framework with Octane (test management tool--They provided support for cucumber BDD)... I have runner class for Octane (https://github.com/HPSoftware/octane-cucumber-jvm/blob/master/src/main/java/com/hpe/alm/octane/OctaneCucumber.java) ... But, we are referring Main class in our project (cucumber.api.cli.Main)... How can I provide referrce to run from OctaneCucumber? or Anyone have done this integration?
*Thread Reply:* https://github.com/trivago/cluecumber-report-plugin
*Thread Reply:* We use JaCoCo, but you can use any coverage tool you like @jradom
*Thread Reply:* Code coverage is independent of test framework and build servers.
*Thread Reply:* Please see if you can use https://github.com/damianszczepanik/cucumber-reporting
*Thread Reply:* Shameless plug 🙂
https://github.com/trivago/cluecumber-report-plugin
*Thread Reply:* I can provide an example of using Cluecumber if not provided by Benjamin already
*Thread Reply:* I have an example project here: https://github.com/trivago/cluecumber-report-plugin/tree/master/example-project
*Thread Reply:* Also, you can check out an example report: http://tech.trivago.com/cluecumber-report-plugin/
*Thread Reply:* I've written this SO post on how to use Cluecumber in case it may be helpful? https://stackoverflow.com/questions/60658416/how-to-use-cluecumber-to-generate-reports?noredirect=1#comment107318396_60658416
*Thread Reply:* Nice! However, I would consider this a short manual how to run Cluecumber standalone if you are not running your tests via Maven. Maybe the question can be edited a bit 🙂
*Thread Reply:* Hi, sorry not sure if I understand the suggestion as my tests are under a maven project?
*Thread Reply:* You wrote "Run your tests and once this is done navigate to the terminal and type mvn cluecumber-report:reporting (exactly as it is)." But this wouldn't be necessary in all cases if you run the tests via Maven and have the reporting included in the lifecycle.
*Thread Reply:* @Benjamin Bischoff we have configured maven / the plugin to generate the report after the integration test, but if no tests have run this fails the build. Is there any way to make it so that doesnt happen?
*Thread Reply:* Hi @mlvandijk, only in case of no tests it fails? What about failing tests?
*Thread Reply:* Then it fails. The report fails if it cannot find the cucumber.json (which happens when no tests have run)
*Thread Reply:* I think you are right, in case of no json inputs, there is a failure. This should be changed.
*Thread Reply:* For us it didn't make a difference as we consider a run without tests a failure as well. But it shouldn't fail because the reporting fails, that's true
*Thread Reply:* I added this to the Cluecumber 2.4.0 roadmap: https://github.com/trivago/cluecumber-report-plugin/issues/240
*Thread Reply:* Currently we are just commenting out the plugin when no tests have been added/are running, but I figured there should be an easier way.
*Thread Reply:* I think the cucumber.json if no tests found is '[]' but I'm not sure
*Thread Reply:* i am pretty sure that no files are generated at all
*Thread Reply:* By the way, there is also the option to skip the report generation completely:
https://github.com/trivago/cluecumber-report-plugin#skip-report-generation
*Thread Reply:* So you can pass the skip property to the plugin to prevent it from running
*Thread Reply:* Cool. I figured there would ve a way but hadnt looked for it yet tbh
*Thread Reply:* I like writing documentation but I also struggle when having to read it 😄
*Thread Reply:* I will try to prepare the next Cluecumber release soon. Just need some more motivation to return.
*Thread Reply:* Please share the details. I am interested. Thanks.
*Thread Reply:* @Taimur Rasheed Sorry about that, I seem to have been distracted for almost a month. Here you go: https://github.com/enkessler/cuke_modeler#projects
Essentially any of those projects is up for grabs but only cuke_linter and cuke_modeler actively need work, as far as I remember. None of it would be difficult work and the task range from easy to ambitious.
*Thread Reply:* Hi, Brian, Perhaps this link coud help you: https://serenity-bdd.github.io/theserenitybook/latest/cucumber.html
As serenity have implemented scrrenplay pattern
*Thread Reply:* yes I've seen serenity / screenplay in several places... although this link to the archetype seems to actually work... thanks!
*Thread Reply:* I personally use Selenide in my projects. Have tried serenity as well, but prefs to Selenide 🙂
*Thread Reply:* @Brian Long there’s a series of blog posts starting here https://cucumber.io/blog/bdd/understandin-screenplay-(part-1)-fundamentals/ by @matt-SmartBear
*Thread Reply:* Serenity is a great place to start if you’re in Java, but just remember that screenplay is a pattern, and you can implement it yourself pretty easily.
*Thread Reply:* ama? unfamiliar. hope it's not terrible. anyway yes we've been working with PageObject and recently I heard a blogcast using words like "training wheels" ... and recommending journey/screenplay. I'm fine with staying with PO, and I'm equally fine with moving to SP if that's where we're going. Just testing the wind.
*Thread Reply:* Those blog posts Steve linked to should give you some idea of how they compare.
*Thread Reply:* I would guess most people use PageObject. Screenplay is not a replacement for PageObject though.
*Thread Reply:* Our test framework uses a ComponentObject pattern which wraps individual component functionalities instead of the whole pages.
*Thread Reply:* Heya, we mostly hang out in #committers and <#C6RLMP3C4|committers-jvm>. You may want to join there.
I don't have the time, expertise or the the interest really to keep the Scala implementation up to date.
So I rather hand this over entirely.
If you want to be maintainer, the tasks mostly come down to:
• Code! Scratch your own itch • Keep dependencies uptodate • Respond to GitHub issues/PRs in a timely manner • Keep the CHANGELOG uptodate • Make releases (Create git tags, update CHANGELOG) • Keep the documentation uptodate • Make the code habitable for new users and contributors • Keep the CI running And of course all that within your own capabilities, interest and availability.
I've created an issue to make it official.
https://github.com/cucumber/cucumber-jvm-scala/issues/21
*Thread Reply:* Ok, thanks for the answer. I'll share it with my company, as we are dependent on it we might want to invest some time to contribute 🙂 In the meantime I'll have a look to the source to understand a bit better how "backends" work.
*Thread Reply:* Can you give an example? I don't think I understand you.
*Thread Reply:* I never use either one of them 🙂 You are free to use whatever makes sense for you 😎
*Thread Reply:* @mpkorstanje it is a simple question. I am asking about verification and validation difference ? actually naming a method in your page class sometimes become very tricky 🙂
*Thread Reply:* @delso yes you can say that but it is only just to get an idea that how different name people are using for their methods
*Thread Reply:* @ichfaisal what is the method doing? You could start it with assert to avoid the validate/verify ambiguity.
*Thread Reply:* They're different words with different meanings. You use them hopefully in their usual meaning. That you ask this question implies that you may see reason to do so otherwise. That at the very least requires some explanation.
*Thread Reply:* I would "verify" an invariant, but "validate" user input. The distinction has to do with one's level of confidence about the possibility of the check failing, how "hard" of an error may be issued if it fails, etc. It's somewhat squishy, but there are certainly cases (such as these examples) where one seems fine to me and the other wouldn't.
*Thread Reply:* Heya,
you are in the right place. You can find the people who commit things in #committers
*Thread Reply:* Are you looking for anything particular to contribute too? A specific language, a tool that you're using, documentation?
*Thread Reply:* Depending on your preferred language, I have several Cucumber related open source projects that could use an additional set of hands. They're not under the official umbrella of Cucumber but it may or may not be easier than jumping into the mono-repos.
*Thread Reply:* Thanks! I know Ruby and am learning Javascript at the moment. Are either of those useful for your projects Eric?
*Thread Reply:* Would this be a good place to start? https://github.com/enkessler/cuke_linter/issues/8
*Thread Reply:* https://rubygems.org/gems/cuke_modeler and https://rubygems.org/gems/cuke_linter are the two most active ones, although any of the Cucumber related ones here could use at least a touch up to make them look vaguely standardized (because every new project started off copy/paste/tweak style but then aged).
*Thread Reply:* That issue would work just fine. It should be simple plug-and-play code and it will get you looking at older solutions to a recurring problem.
*Thread Reply:* "required tag starts with" and "unique scenario names" may be the easiest of the ones left.
*Thread Reply:* Feel free to DM me on Slack or in the GitHub issues themselves if you want to give it a go and have questions.
*Thread Reply:* Thank you! I'll take a look at the two you suggested and see how I get on. 🙂
*Thread Reply:* <#C5YHPPJMP|help-cucumber-jvm> is a more appropriate channel for that question.